Page 1
A Structure Function Model for 1-Pyrroline-5’ Carboxylate Reductase
Shukla,V.B., Johnson, C.A and Hawes, J.W.
Department of Chemistry and Biochemistry
Miami University
Oxford, Ohio 45056
Phone: 513-529-8072
FAX: 513-529-5715
[email protected]
Running Title- 1-Pyrroline-5’ Carboxylate Reductase
Page 2
Abstract
We present a structure / function model for 1-pyrroline-5’ carboxylate reductase (P5CR)
based on enzymatic and structural similarity to the β-hydroxyacid dehydrogenase enzyme
family. P5CR is similar to the β-hydroxyacid dehydrogenases in that it is non-metal-dependant
and catalyzes redox chemistry with specific carboxylate substrates. Recombinant E. coli P5CR
displays circular dichroism spectra similar to that of β-hydroxyisobutyrate dehydrogenase,
suggesting similarities in secondary structural contents. P5CR amino acid sequences from
diverse species display homology in the regions corresponding to the known functional domains
of the β-hydroxyacid dehydrogenases. These include the N-terminal dinucleotide cofactor-
binding domain, the carboxylate substrate-binding domain, and the catalytic domain. Site-
directed mutagenesis of a conserved threonine residue in the proposed carboxylate-binding
domain of the E. coli P5CR produces mutant enzymes with reduced catalytic efficiency.
Treatment of the recombinant E. coli P5CR with either iodoacetamide or tetranitromethane
results in inactivation of the enzyme. Quantitation of nitrotyrosine residues shows that
inactivation occurs concomitant with the production of two moles of nitrotyrosine per mole of
enzyme. NADP+ protects the enzyme from inactivation with iodoacetamide but enhances
inactivation with tetranitromethane. E. coli P5CR has three tyrosine residues in its amino acid
sequence, all of which lie within the proposed functional domains. Semi-conservative
replacement of each of these tyrosine residues with phenylalanine results in distinct kinetic
effects, including increased catalytic efficiency. Neither proline nor a variety of β-hydroxyacids
inhibit P5CR; however, glutamate inhibits the enzyme at sub-millimolar concentrations. Based
on this model we propose that γ-glutamate semialdehyde (the straight-chain isomer of pyrroline
carboxylate) may actually be the true substrate for P5CR.
Key Words: Active site, Pyrroline-5-carboxylate, NADPH, proline, oxidoreductase, site directed
mutagensis, , L-thiazolidine-4-carboxylate, glutamate.
Page 3
Introduction
Pyrroline-5’carboxylate (P5C) is an intermediate in the biosynthesis of the amino acid
proline, and is situated at a branch point between the citric acid cycle and the urea cycle through
conversion to ornithine. As a redox couple, proline and P5C are reported to form a cycle
capable of transferring reducing equivalents into mitochondria as proline and oxidizing potential
out as P5C (1). This cycle depends on the presence of proline oxidase in mitochondria and
pyrroline carboxylate reductase in the cytosol. P5C, as a precursor of proline, has multiple
physiological roles in plants (heavy metal and osmotic stress), bacteria, and animal tissues. (2-
10).
Pyrroline-5’-carboxylate reductase (P5CR; EC 1.5.1.2) is the enzyme which catalyzes the
final reaction in proline biosynthesis by converting P5C to proline with either NADH or NADPH
as cofactor. It may also regulate metabolism limited by NADP+ concentrations such as the
pentose phosphate pathway (11). P5CR is unusual compared to other oxidoredutases as it acts
unidirectionally with its natural substrate, catalyzing only the reduction of P5C and not the
oxidation of proline, whereas most oxidoreductase-catalyzed reactions are fully reversible. It
does, however, oxidize a proline analogue, L-thiazolidine-4-carboxylate. How this occurs,
mechanistically, is not known. Bacterial proline synthesis from glutamate occurs via three
enzymatic reactions catalyzed by -glutamyl kinase (proB gene product, EC 2.7.2.11), -glutamyl
phosphate reductase (proA gene product, EC 1.2.1.41) and P5CR (proC gene product). The E.
coli proC gene was shown to encode a pyrroline carboxylate reductase (4). For the majority of
bacteria the proB and proA genes constitute an operon which is distant from proC on the
chromosome (8). There are reports that synthesis of P5CR in E. coli is not subject to repression
by proline (12, 13). Rossi et. al reported that synthesis of P5CR is not repressed by growth in the
presence of proline, but it is inhibited by only high levels of the reaction end products, proline
and NADP+ (14). Deutch et.al. have demonstrated the co-purification of L-thiazolidine-4-
caboxylate dehydrogenase activity and P5CR activity, although proline was not oxidized to P5C
by this partially purified protein in the presence of NADP+ (15). Rossi et. al. also reported that
partially purified P5CR from E. coli catalysed conversion of P5C to proline but not the reverse
reaction (14). Fujii et al. demonstrated that P5CR also catalyzes the reduction of 1-piperideine-
6-carboxylate to L-pipecolic acid (16). However, oxidation of pipecolic acid was not analyzed.
Kenklies et. al. reported the purification and sequence of P5CR from Clostridium sticklandii, and
Page 4
its expression (17). Deutch et. al. (18), and Rossi et. al. (14) reported the E. coli P5CR gene
sequence, protein over-expression, and purification of the enzyme. P5CR has also been purified
approximately 200 fold from crude extract of baker’s yeast (19). With this purified enzyme the
Km value for DL-P5C was 0.8 X 10-4; for NADH it was 4.8 X 10-5 and for NADPH was 5.6
X10-5. Several reports suggest that mammals express two distinct pyrroline carboxylate
reductases with differences in cofactor specificity (20). Examination of mammalian genome
databases clearly show at least two homologues, although no particular sequence elements in
these homologous sequences provide clear clues as to residues that might affect cofactor
specificity. No studies have focused on structure / function relationships in this enzyme from
any species; nor have any previous studies suggested any possible models for structure / function
relationships.
We report a unique structure / function model of P5CR based on structural and enzymatic
comparisons to the β-hydroxyacid dehydrogenases. The β-hydroxyacid dehydrogenases are a
family of enzymes with characteristic functional domain structure and substrate specificities for
various β-hydroxycarboxylic acids (21, 22). Some members of this family (such as tartronate
semialdehyde reductase) function physiologically in the reductive direction (22). We show that,
although there is limited overall amino acid sequence homology among these proteins, there is
conservation of critical structural features throughout the known functional domains of the β-
hydroxyacid dehydrogenases. Using recombinant E. coli P5CR, we have tested this model
through the application of specific chemical modifications and site-directed mutagenesis.
Results of specific mutations are consistent with a model based on structural and functional
similarity between β-hydroxyacid dehydrogenases and pyrroline-carboxylate reductase.
Inhibition of E. coli P5CR by glutamate, but not by proline suggests that γ-gluatamate
semialdehyde (the straight-chain isomer of pyrroline carboxylate) may actually be the true
substrate for P5CR.
Materials and Methods
Strains, media and growth conditions
Page 5
E. coli DH10B was used in this study. Cultures were grown at 37◦C in TY broth
containing per liter: 10 g tryptone, 5 g yeast extract, and 0.5 g NaCl. Cultures were maintained
on solid medium (1.5% agar). Ampicillin (50 µg per ml) was included as appropriate for
recombinant cultures.
Cloning of Pyrroline-5’-Carboxylate Reductase (proC) gene
For this purpose PCR amplification of the E. coli proC gene was performed using
primers ECP5CR5’ (ATGGAAAAGAAAATCGGTTTTATT) and ECP5CR3’
(TCAGGATTTGCTG AGTTTTTCTG). High molecular weight E. coli genomic DNA (Sigma)
was used as template. The Biorad Icycler was used for PCR with one cycle at 95°C for 2 min, 45
cycles of: 95°C 33sec, 48°C 60sec, and 70°C 1.3min. The PCR product size was confirmed by
gel electrophoresis using 1 % agarose. After purification by alcohol precipitation, the PCR
product was cloned into pTrcHis-TOPO vector using pTrcHis TOPO TA Expression kit
(Invitrogen Life Technologies) and transformed into chemically competent E.coli DH10B. The
recombinant clones were selected with 50µg ml−1 ampicillin. Plasmids were purified using
Perfect Prep plasmid mini kit (Eppendorf), digested using BamH1 and EcoR1, and analyzed by
agarose gel electrophoresis to confirm the size of the cDNA insert. The orientation of the insert
was confirmed by PCR using pTrcHis Forward Primer (GAGGTATATATATTAATGTATCG)
and ECP5CR5’ primer (described above). The clones were sequenced using Dyenamic ET
terminator (Amersham Bioscience) and an Applied Biosystems 310 DNA Sequencer. Primers
used for sequence analysis were TOPO Express Forward (TATGGCTAGCATGACTGGT) and
T122SEQ (TGCTTAGCGAAATCACCTC).
Site-directed mutagenesis
For this purpose the QuickChange Site-Directed Mutagenesis Kit (Stratagene) was used.
The primers used for Y122 mutation were CGGAAAATTATCCGCGCCATGCCGAACGCTCC
CGCACTGGTTAATGCCGGGATGACC and GTCATCCCGGCATTAACCAGTGCGGGAGC
GTTCGGCATGGCGCGGATAATTTTCCG. For Y122S mutation, primers used were CGGA
AAATTATCCGCGCCATGCCGAACTCTCCCGCACTGGTTAATGCCGGGATGACC and G
GTCATCCCGGCATTAACCAGTGCGGGAGAGTTCGGCATGGCGCGGATAATTTTCCG.
The primers for Y35F mutation were GCAAATCTGGGTATTCACCCCCTCCCCG and
CGGGGAGGGGGTGAATACCCAGATTTGC. The primers for Y178F mutation were CGGT
Page 6
TCTTCGCCAGCCTTCGTATTTATGTTTATCGA and TCGATAAACATAAATACGAAGG
CTGGCGAAGAACCG. The primers used for Y201F mutation wereCGCGCCCAGGCG TTT
AAATTTGCCGCTC and GAGCGGCAAATTTAAACGCCTGGGCGCG. The mutagenic PCR
was performed using purified plasmid DNA (from a verified clone of P5CR) as template with the
addition of 8% DMSO. The condition for mutagenic PCR was one cycle at 95°C for 30 sec, 16
cycles of : 95°C 30sec, 55°C 60sec, and 68°C 11min. The mutations were confirmed by
sequence analysis using either TOPO Express Forward primer or T122SEQ primer.
Recombinant Enzyme Purification
For expression of recombinant enzymes, cultures were grown at 37◦C in TY broth with
ampicillin (50 µg ml−1). When OD600 was reached between 0.5 – 0.8 (estimated using Ultraspec
2100 pro UV/visible spectrophotometer, Amershan Bioscience), 0.5 mM IPTG was then added
to induce production of the wild-type or mutant recombinant enzymes. After 24 hours of
induction, cell mass was harvested by centrifugation at 6000 X g, at 4◦C, using an RC5C
centrifuge (Sorvall Instrument) and was resuspended in 15 ml of 25 mM Tris-HCl buffer (pH
7.0) with 0.5% Tween-20. The suspension was sonicated 5 times on ice for 30 Sec with
intermittent gaps of 30 sec each at continuous mode using a Sonic Dismembrator (model 100
Fisher Scientific). The suspension was centrifuged at 12,000 X g, at 4◦C, for 20 min to remove
cell debris using an Eppendorf centrifuge 5810. Supernatant was mixed in equal quantities with
2X column buffer. The column buffer had composition 50mM Na2HPO4, 10mM 2-
mercaptoethanol, 250 mM NaCl, 5mM imidazole, pH 8.0. This solution was passed through 10
ml Ni-NTA Agarose beads (Qiagen) equilibrated in the same buffer. The columns were then
eluted with 30 ml of column buffer containing 20, 40, 60, and 200 mM imidazole. Protein
concentrations of all the 200mM imidazole fractions were analyzed using Bradford reagent
(BioRad), using BSA as a protein standard. The samples with detectable levels of proteins were
pooled and frozen in aliquots with 10% glycerol.
SDS –PAGE Analysis
For gel electrophoresis 25µl of the samples along with the protein markers were loaded
on 10% SDS-PAGE gels (23). After Commossie blue staining the gels were imaged using the
Versa Doc 3000 imaging system (Biorad). The corresponding bands of protein were excised
after successive washing and subjected to in-gel digestion with sequence grade trypsin in 10mM
Page 7
ammonium bicarbonate (pH 9.0). The digested peptides were extracted and concentrated, and
suspended in 10% acetonitrite with 0.05% TFA. This was mixed in equal volume with 4-
hydroxy cinnamic acid (10mg/ml) in 50% acetonitrile with 0.05% TFA. MALDI-TOF mass
spectra were collected using a Brucker reflex mass spectrometer in reflectron mode.
Enzyme kinetic analysis
The enzyme activity was measured according to the method published by Deutch et. al
(2001) at 37°C measuring an increase in absorbance at 340 nm in a Cary 1E UV-Visible
spectrophotometer (Varian). Reaction mixtures normally contained 300 mM Tris-HCl, pH 7.2,
0 to 12.5 mM L-thiazolidine-4-carboxylate , and 0 to 50 µM NADP+ in a total volume of 1.0 ml.
Activities were calculated in mmoles of NADPH formed per min from the first 30 s of 2-min
assays, using a mM extinction coefficient of 6.22 for NADPH. For proline oxidation assays 0 to
10 mM concentration of L-proline were used as substrate in place of L-thiazolidine-4-
carboxylate.
Chemical Modifications and Inhibition
For treatment with iodoacetamide, aliquots of enzyme stored with 5 mM 2-
mercaptoethanol were diluted 5-fold in Tris-HCl buffer, pH 7.2 and treated with 5 mM
iodoacetamide from freshly prepared stock solutions made in the same buffer. P5CR activity
was then measured immediately, following the given incubation times, using L-thiazolidine-4-
carboxylate as substrate exactly as described above. For treatment with tetranitromethane
(TNM), aliquots of enzyme stored with 5 mM 2-mercaptoethanol were diluted in Tris-HCl
buffer, pH 7.2 and treated with 83 mM TNM from freshly prepared stock solutions in 95%
ethanol. Control reactions were prepared with equal volumes of ethanol alone. P5CR activity
was then measured immediately, following the given incubation times, using L-thiazolidine-4-
carboxylate exactly as described above. For measurement of the stoichiometry of nitrotyrosine
formation, frozen aliquots of wild-type or mutant enzyme were desalted by passage through pre-
packed PD10 gel filtration columns (Pharmacia) and treated with TNM exactly as described
above. Production of nitrotyrosine was monitored throughout the course of these reactions by
measurement of OD428. Moles of nitrotyrosine produced were calculated using a molar
extinction coefficient of 4100 as reported by Sokolovsky et al (24). Treatment with
phenylmethylsulfonyl fluoride (PMSF) was performed exactly as described above for TNM
Page 8
accept for using PMSF stock solutions freshly prepared in 95% isopropanol. In this case,
isopropanol was added to control reactions.
CD Spectrapolarimetry
Circular dichroism spectra were recorded at 25°C using a Jasco J-810 spectrapolarimeter.
Spectra were recorded with a cell path length of 0.1 cm and a wavelength range of 300–200 nm.
Each spectrum was averaged from five separate recordings. Immediately before measuring the
spectra, protein samples were desalted using PD10 columns exatly as described above. 100 µg
of protein were used for each measurement. Protein was quantitated using the Bradford protein
assay (Biorad) with BSA employed as standard.
Structural Modeling
Structural modeling of enzyme substrates was performed using HyperChem software
(Hypercube, Inc.). Geometry optimization was performed using Polak-Ribiere (conjugate
gradient) method.
Results and Discussion
Comparison of Substrates and Reaction Characteristics
The reactions catalyzed by P5CR and those catalyzed by β-hydroxyacid dehydrogenases
are different but have certain features in common. P5CR does not catalyze the oxidation of a β-
hydroxyacid. It does however, catalyze the NAD(P)+-dependent oxidation / reduction of a
carbon distal to a carboxylate, which is structurally and mechanistically similar to reactions
catalyzed by known β-hydroxyacid dehydrogenases. P5CR also has enzymatic similarity to the
β-hydroxyacid dehydrogenases in that it is non-metal-dependant. A major difference between
P5CR and the β-hydroxyacid dehydrogenases is the fact that P5CR is unidirectional with its
natural substrate (P5C) and does not oxidize proline. It does, however, oxidize the substrate
analogue L-thiazolidine-4-carboxylate (thiaproline). It is not clear what structural features
provide for this difference in reaction characteristics between proline and thiaproline. We have
modeled these substrates and compared torsional bond angles in the most stable structures after
performing energy minimizing geometry optimizations (Fig 1). This analysis indicates that the
pyrroline carboxylate ring is stabilized by a significantly different conformation than the ring
Page 9
structures of proline or thiaproline (thiazolidine-4-carboxylate). A major structural difference
between pyrroline carboxylate and proline is the position of the carboxylate atoms relative to the
nitrogen atom in the ring, reflected in the torsional bond angles between these atoms (Fig 1).
Comparison of the torsional bond angles between the carboxylate oxygen atoms and the ring
nitrogen show that the pyrroline and proline rings take on significantly different conformations
relative to the plane of the ring and the carboxylate (Fig 1). Torsional angles between C2 and the
C5 protons also reflect this difference between the pyrroline and proline rings. Torsional angles
between the carboxylate carbon and the C5 of the pyrroline, proline, and thiazolidine rings show
dramatic conformational differences between all three compounds. These structural differences
may explain, in part, the irreversible nature of the enzyme-catalyzed reduction of pyrroline
carboxylate, either through differences in binding of these compounds to the enzyme active site
or through structural inability of the enzyme to contact the appropriate proton in the proline ring
for hydride transfer to NADP+.
P5CR should require chemistry for carboxylate-binding and for acid/base catalysis
similar to the chemistry required of the β-hydroxyacid dehydrogenases. The features of the
P5CR-catalyzed reaction in common with that of the β-hydroxyacid dehydrogenases are
carboxylate-binding, dinucleotide cofactor-binding, acid/base catalysis and hydride transfer to
the cofactor. Judging based on the chemistry alone, what is likely most different is the nature of
the acid/base catalyst. Although the P5CR-catalyzed reaction was not previously believed to
involve an acid/base-catalyzed proton abstraction from a carbonyl group, our model indicates
that the architectural frame-work of this enzyme is likely similar to a family of enzymes that bind
specifically to β-hydroxyacids and semialdehydes. This model suggests the possibility that the
true substrate of P5CR may actually be γ-glutamate semialdehyde, an isomeric form of pyrroline
carboxylate (Fig 2). Previous studies related to proline biosynthesis assumed that pyrroline
carboxylate is the direct substrate for this enzyme and that it is produced in solution
spontaneously from γ-glutamate semialdehyde. We propose that γ-glutamate semialdehyde may
be the true substrate for this enzyme and the enzyme may facilitate conversion to pyrroline
carboxylate (Fig 2), which may be more efficient, metabolically, than the spontaneous
conversion of the semialdehyde to pyrroline carboxylate. Considering this possibility, we tested
the inhibition of P5CR by glutamate, proline, and β−hydroxyacids. Glutamate was indeed found
Page 10
to be a much more potent inhibitor of purified recombinant P5CR than any other acids
previously tested. Gluatamate at concentrations between 250 µM and 1mM produced significant
inhibition in the NADP+-dependent oxidation of thiaproline whereas neither proline nor a variety
of other common acids inhibited the enzyme even at concentrations over 10 mM (Fig 3). Kinetic
analysis of this inhibition clearly showed that glutamate is not a competitive inhibitor of
thiaproline, but inhibits through a mixed uncompetitive mechanism involving changes in both
Vmax and Km. It is particularly notable that proline does not inhibit the enzyme even at high
concentrations. This level of inhibition and specificity is consistent with an enzyme-assisted
(non cofactor-dependent) conversion of glutamate-semialdehyde to pyrroline carboxylate
followed by the NADPH-dependent reduction to proline (Fig 2). This model, including the
structural differences in substrates shown in Figure 1, may explain, in part, the inability of the
enzyme to catalyze the oxidation of proline. Alternatively, glutamate might have sufficient free
rotation to be able to bind to an enzyme active site that ordinarily binds pyrroline carboxylate.
However, such this alternative might be expected to result in competitive inhibition.
Amino Acid Sequence Homology
Comparison of primary structures of P5CR and well-established β-hydroxyacid
dehydrogenases show only a low level of overall identity. Figure 4 shows an amino acid
sequence alignment of P5CR sequences from diverse species with those of several well-
established β-hydroxyacid dehydrogenases. Many of the glycine and alanine residues conserved
in the β-hydroxyacid dehydrogenases, as well as several highly conserved proline residues, are
also conserved in P5CR, probably demonstrating a common structural framework. Closer
inspection predicts putative structural elements that reflect a possible shared functional domain
structure. One of the most obvious and striking similarities between the pyrroline carboxylate
reductases and the β-hydroxyacid dehydrogenases is in the N-terminal dinucleotide cofactor-
binding domain, which is highly characteristic of this family. This domain consists of a motif
with a sequence of VGFIGXGXMGXXXAX5AG. The sequence conservation in the P5CRs fits
extremely well to this consensus sequence including the distribution of hydrophobic residues, the
conserved methionine residue, and the most highly conserved alanine and glycine residues (21).
The next major domain is the carboxylate-binding domain downstream from the cofactor-
Page 11
binding domain. The homology of the P5CRs with the previously published consensus sequence
for this domain is less than that of the β-hydroxyacid dehydrogenases alone (21, 22). The beta-
hydroxyacid dehydrogenases display a highly conserved substrate-binding motif with a sequence
of DAPVSGGXXXA (21, 22), followed by a number of well-conserved glycine and alanine
residues. Upstream from this motif is a highly conserved serine residue directly preceded by a
stretch of considerably hydrophobic but otherwise non-conserved sequence. The P5CR
sequences share a number of important features in this domain. As with the N-terminus, most of
the conserved residues are glycines and alanines that probably reflect a similar structural
framework (Fig 4). The conserved serine residue (in the motif DAPVSGG) believed to be
functional in carboxylate-binding in the β-hydroxyacid dehydrogenases (21,22) is semi-
conserved as a threonine residue in the P5CR sequences and is invariant in the P5CRs (Fig 4).
This highly conserved residue was shown to make direct contact with a substrate carboxylate
oxygen atom in the crystal structure of 6-phosphogluconate dehydrogenase (25). Furthermore,
both of these conserved serine and threonine residues are preceded by a highly conserved proline
residue in either a PXS or PXT sequence (Fig 4). We predict that this conserved threonine
residue in P5CR may be specifically important to carboxylate-binding in P5CRs, and that this
motif further defines a conserved structure for carboxylate-binding among all of these enzymes.
This is not surprising given that all the β-hydroxyacid dehydrogenases, as well as P5CR, must
bind a carboxylate substrate in their active sites (non-acid analogues are not active substrates).
The catalytic domain is clearly conserved in structure, but differs between the β-
hydroxyacid dehydrogenases and P5CRs in several respects. The putative catalytic lysine
residue present in the β-hydroxyacid dehydrogenases (Fig 4) is absent in the P5CR sequences,
whereas other features of this domain are well conserved. Perhaps it should not be surprising
that this conserved catalytic residue in all known β-hydroxyacid dehydrogenases is absent in
P5CRs or that its function could be performed by another residue. There may be specific
residues to support the reduction of P5C, a reaction similar to (but somewhat different from) β-
hydroxyacid oxidation. However, these catalytic residues appear to be situated in a region with a
similar structural framework. There is a highly conserved GXXGXG consensus sequence in this
region, although it is shifted downstream slightly in the P5CR sequences (Fig 4). This motif
likely represents a conserved structural element such as a turn or bend in secondary structure that
may function in positioning of catalytic residues. The consensus sequence downstream from
Page 12
this, containing highly conserved glutamate, alanine, and leucine residues, is highly conserved in
the P5CRs. A number of amino acid residues in this domain could possibly serve as specialized
acid/base catalysts for P5CR, and our hypothesis predicts that the most important catalytic
residues for P5CR will reside in this region, downstream from the GXXGXG motif.
This hypothetical model predicts a number of amino acid residues and domains that are
likely of structural and catalytic importance in P5CR, which would not have been predicted
without this structural comparison. These include the very distinct N-terminal dinucleotide
cofactor-binding domain, a carboxylate-binding domain with a conserved threonine residue, and
a catalytic domain with a number of conserved glycine, alanine, and proline residues. This
comparison now provides a model for further structure / function studies of P5CRs by site-
directed mutagenesis and chemical modifications.
Expression and Purification of Recombinant E. coli P5CR
The proC gene of E. coli was amplified by PCR using primers based on the known DNA
sequence and using purified E. coli genomic DNA as the template for PCR. This cDNA was
ligated in frame to the pTrcHisTOPO vector (Invitrogen) for over-expression in E. coli cultures.
Induction with IPTG led to the appearance of a largely over-expressed band of protein
approximately 30 kD in size. This protein was purified by nickel affinity chromatography as
described in Materials and Methods. The resulting protein was over 90% pure as estimated by
SDS-PAGE analysis and was verified as E. coli P5CR by trypsin digestion and MALDI-TOF
mass spectral analysis of the resulting peptides. This purified, wild-type enzyme was not active
in the oxidation of proline with either NADP+ or NAD+ as cofactor, but oxidized thiaproline in
the presence of NADP+ with a specific activity similar to that previously reported for the native
enzyme (Table 1) (15). Two mutations of a conserved threonine residue were produced as well
as three mutations of tyrosine residues. Each of these mutants were expressed at similar levels to
wild-type enzyme and each resulted in identical levels of purity when purified by nickel affinity
chromatography. The mutant enzymes were also analyzed by CD spectrapolarimetry and
displayed spectra identical to that of wild-type enzyme. These purified enzymes were found to
be stable for at least one month when stored at -20°C in the presence of β-mercaptoethanol and
glycerol. In the absence of β-mercaptoethanol the purified enzymes were only stable for several
days.
Page 13
CD Spectral Comparison of P5CR and Hydroxyisobutyrate Dehydrogenase.
The sequence homology described above predicts a similar structural architecture
between P5CRs and β−hydroxyacid dehydrogenases, despite a relatively low overall sequence
identity. To test this hypothesis CD spectrapolarimetry was used to compare secondary
structural contents of E. coli P5CR with that of rat β-hydroxyisobutyrate dehydrogenase, one of
the most well-characterized β-hydroxyacid dehydrogenases. Examination of the CD spectra for
these two enzymes indicate that they have very similar secondary structures (Fig 5). Comparison
of the ellipticities at 208nm and 220nm suggest very similar content of α-helices (26).
Role of Conserved Threonine Residue
As described above, the sequence homology within the previously proposed carboxylate
substrate-binding domain suggests a role for a conserved threonine residue in P5CR. This
residue (T122 in the E. coli enzyme) is absolutely invariant in all known P5CR sequences, also
suggesting the possibility of a conserved functional role. To test whether this threonine residue
may be catalytically important, we performed site-directed mutagenesis substituting it with either
alanine or serine. As shown in Table 1, these substitutions both produced increased Km values
for thiaproline as well as lowered Kcat values, resulting in significantly reduced catalytic
efficiencies. Alanine substitution produced a greater increase in the Km for thiaproline
(approximately 10-fold) compared to serine substitution. Serine substitution also produced a
much less dramatic change in the Vmax. Both the alanine and serine substitutions led to changes
in the Km for NADP+; however, the serine substitution produced a greater increase in this
parameter than the alanine substitution. These results are consistent with a proposed role for
substrate binding by this conserved threonine residue.
Chemical Modification of Ser, Tyr and Cys Residues
To begin determining other possible active site residues, recombinant P5CR was tested
for inhibitory effects of a number of amino acid side chain-specific chemical modifiers.
Although there is a highly conserved serine residue at the N-terminal end of the substrate-
Page 14
binding domain (Fig 4) there is no affect of PMSF even at high concentrations. This suggests
that there are no binding events or catalytic events involving acid/base chemistry with this
conserved serine residue or any other serine residue. PMSF usually reacts with specialized
serine residues with pKa values lowered by the structures within their local environment. This
must not be the case for the conserved serine residue in the β-hydroxyacid dehydrogenases and
P5CRs.
Treatment with iodoacetamide resulted in over 90% inactivation of recombinant E. coli
P5CR, with almost complete protection provided by the addition of NADP+ (Fig 6). This is
similar to the effects of cysteine modifiers on β-hydroxyacid dehydrogenases; however, there are
no specifically conserved cysteine residues in this family of enzymes, and none believed to be
involved in catalysis or substrate binding (21, 22, 27). Nevertheless, most of the β-hydroxyacid
dehydrogenases are sensitive to a variety of chemical modifiers of cysteine residues, and this
appears to be true of P5CR as well. There are no conserved cysteine residues in the amino acid
sequences for P5CR, but there is a unique cysteine residue directly within the proposed N-
terminal cofactor-binding domain of the E. coli P5CR sequence (Fig 4).
Treatment with tetranitromethane (TNM) also resulted in inactivation of E. coli P5CR.
Unlike the case with iodoacetamide, the presence of NADP+ did not protect the enzyme from
inactivation, but actually enhanced the inactivation with TNM (Fig 6). Measurement of
nitrotyrosine production by absorbance at 428 nm revealed that inactivation of wild-type E. coli
P5CR occurred concomitant with the production of 2 moles nitrotyrosine per mole of enzyme. It
is possible that cofactor alone cannot afford protection against TNM inactivation if inactivation
occurs through the labeling of multiple, functionally unrelated sites. These results are consistent
with the presence of tyrosine residues in the active site of E. coli P5CR. Indeed, there are only
three tyrosine residues in the E. coli P5CR sequence, and each of these residues lie within or
very near to the proposed functional domains shown in Figure 4. One of these tyrosine residues
(Y35 in the E. coli enzyme) lies in the N-terminal cofactor-binding domain. The other two
tyrosine resides lie directly within the proposed catalytic domain. One of these tyrosine residues
(Y178 in the E. coli enzyme) occurs just downstream from the characteristic GXXGXG motif in
the proposed catalytic domain and is completely conserved in the P5CR sequences from widely
variant species.
Page 15
Site-directed mutagenesis of Tyrosine residues
The inactivation by TNM, the stoichiometry of nitrotyrosine production, and the presence
of tyrosine residues in the proposed functional domains of E. coli P5CR strongly suggest the
presence of active site tyrosine residues in this enzyme. To test the catalytic importance of these
residues, replacement with phenylalanine residues was performed by site-directed mutagenesis.
Surprisingly, each of these very semiconservative replacements resulted in increases in Vmax
and Kcat (Table 1). Although consistent with active site locations, it is not exactly clear why
these replacements should result in increased catalytic rate. Phenylalanine replacement of Y36
(located in the proposed cofactor-binding domain) resulted in relatively small changes in the Km
for either NADP+ or thiaproline, but produced approximately 5-fold increase in Vmax.
Phenylalanine replacement of Y178, the invariant tyrosine, resulted in relatively small changes in
Km values and also a smaller (2.7-fold) increase in Vmax compared to the Y36 mutation. This
result rules out an identity for this residue as a possible acid/base catalyst. Phenylalanine
replacement of Y201, also in the proposed catalytic domain, resulted in the most dramatic
changes, with a nearly 5-fold increase in the Km for NADP+, a 3-fold increase in the Km for
thiaproline, and a 6-fold increase in Vmax. It is not clear why this mutation should result in an
increased catalytic rate. It is possible that this structural change increases the rate-limiting step
in this reaction, which is presumably the release of reduced cofactor. Increases in the Km for
NADP+ cofactor may be consistent with this explanation. It is also possible that these surprising
increases in catalytic rate are related to an increased efficiency in binding of the unnatural
substrate analogue used in the assay, thiaproline. However, the observed increases in Km for
thiaproline, though small, are not consistent with this explanation.
Acknowledgements
We thank Dr. Mike Crowder and Dr. Rich Taylor (Miami University) for helpful discussion.
Page 16
References
1. Hagedorn, C. H. and Phang, J. M. (1983) Transfer of Reducing Equivalents into
Mitochondria by the interconversion of Proline and ∆1-Pyrroline-5-carboxylate. Arch.
Biochem Biophys. 225, 95-101.
2. Delauney A.J. and Verma D.P.S. (1993) Proline biosynthesis and osmoregulation in
plants. Plant J. 4, 215-223.
3. Baumberg, S. and. Klingel U. (1993). Biosynthesis of arginine, proline, and related
compounds, p. 299-306. In A. L. Sonenshine, J. A. Hock, and R. Losick (ed.), Bacillus
subtilis and other Gram-positive bacteria: biochemistry, physiology, and molecular
Genetics. American Society for Microbiology, Washington, D. C.
4. Leisinger, T. (1996) Biosynthesis of proline, p 434-441. In F.C. Neidhardt et al (ed),
Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed. American
Society for Microbiology, Washington, D.C.
5. Csonka, L. N. (1989) Physiological and genetic responses of bacteria to osmotic stress.
Microbiol. Rev. 53, 121-147.
6. Csonka, L. N. and Hanson A. D. (1991). Prokaryotic osmoregulation: genetics and
physiology. Annu. Rev. Microbiol. 45, 569-606
7. Kosuge, T. and Hoshino T. (1998) Construction of a proline-producing mutant of the
extremely thermophilic eubacterium Thermus thermophilus HB27. Appl. Environ.
Microbiol. 64, 4328-4332.
8. Sleator R. D., Gahan C. G. M., and Hil C. (2001) Mutations in the Listerial proB Gene
Leading to Proline Overproduction: Effects on Salt Tolerance and Murine Infection
Applied Environ. Microbiol. 67, 4560-4565.
9. Culham, D. E., Dalgado C., Gyles C. L., Mamelak D., MacLellan S., and Wood J. M.
(1998) Osmoregulatory transporter ProP influences colonization of the urinary tract by
Escherichia coli. Microbiology 144, 91-102.
10. Schwan W. R., Coulter S. N., Ng E. Y. W., Langhorne M. H., Ritchie H. D., Brody L. L.,
Westbrock-Wadman S., Bayer A. S., Folger K. R., and Stover C. K. (1998). Identification
Page 17
and characterization of the PutP proline permease that contributes to in vivo survival of
Staphylococcus aureus in animal models. Infect. Immun. 66, 567-572.
11. Yeh G C., Phang J. M. (1988) Stimulation of phosphoribosyl pyrophosphate and purine
nucleotide production of pyrroline 5-carboxylate in human erythrocytes. J. Biol. Chem,
263, 13083-13089.
12. Baich A. and Pierson, D. J. (1965) Control of proline synthesis in Escherichia coli.
Biochim. Biophys. Acta. 104, 397-404.
13. Condamine H. (1971) Sur la regulation de la production de proline chez E. coli K-12.
Ann Inst. Pasteur (Paris) 120, 126-143.
14. Rossi J. J., Vender J., Berg, C. M., and Coleman W. H. (1977) Partial purification and
some properties of delta1-pyrroline-5-carboxylate reductase from Escherichia coli. J.
Bacteriol. 129, 108-114.
15. Deutch C. E., Klarstrom J. L., Link, C. L., and Ricciardi D. L. (2001) Oxidation of l-
Thiazolidine-4-Carboxylate by ∆1-Pyrroline-5-Carboxylate Reductase in Escherichia
coli. Curr. Microbiol. 42, 442-446.
16. Fujii T., Mukaihara M., Agematu H., Tsunekawa H. (2002) Biotransformation of L-
lysine to L-pipecolic acid catalyzed by L-lysine 6-aminotransferase and pyrroline-5-
carboxylate reductase. Biosci. Biotechnol. Biochem. 66, 622-627.
17. Kenklies J., Ziehn, R., Fritsche K., Pich, A., Andreesen J.R. (1999) Microbiol. Proline
biosynthesis from L-Ornithine in Clostridium sticklandii: purification of DELTA1-
pyrroline-5-carboxylate reductase, and sequence and expression of the encoding gene,
proC, 145, 819-826.
18. Deutch A.H., Smith, C. J., Rushlow K.E. Kretschmer P.J. (1982) Escherichia coli delta 1-
pyrroline-5-carboxylate reductase: gene sequence, protein overproduction and
purification. Nuc. Acids Res. 10, 7701-7714.
19. Matsuzawa T., and Ishiguro I. (1980) Delta-1-Pyrroline-5-caboxylate reductase from
baker’s yeast. Purification, properties and its application in the assays of L-delta-1-
pyrroline-5-caboxylate and L-Ornithine in tissue, Biochim. Biophysica Acta. 613, 318-
323.
Page 18
20. Lorans, G., and Phang, J.M. (1981) Proline synthesis and redox regulation: differential
functions of pyrroline-5-carboxylate reductase in human lymphoblastoid cell lines.
Biochem Biophys Res Commun. 101, 1018-25.
21. Hawes J.W., Harper E.T., Crabb D.W., Harris R.A. (1996) Structural and mechanistic
similarities of 6-phosphogluconate and 3-hydroxyisobutyrate dehydrogenases reveal a
new enzyme family, the 3-hydroxyacid dehydrogenases. FEBS Lett 389, 263-267.
22. Njau R.K., Herndon C. A, Hawes J. W. (2000) Novel beta-hydroxyacid dehydrogenases
in Escherichia coli and Haemophilus influenzae. J. Biol Chem. 275:38780-38786.
23. Laemmli U.K. (1970) Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature 227, 680-685.
24. Sokolovsky M., Riordan J. F. and Vallee B.L. (1966) Tetranitromethane. A reagent for
nitration of Tyrosyl residues in protein. Biochemistry 5, 3582-3589.
25. Adams, M.J., Ellis, G.H., Gover, S., Naylor, C.E., and Phillips, C. (1994)
Crystallographic study of coenzyme,coenzyme analogue and substrate binding in 6-
phosphogluconate dehydrogenase: Implications for NADP specificity and the enzyme
mechanism. Structure, 2, 651–668.
26. Yang, J.T., Wu, C.S., and Martinez, H. M. (1986) Calculation of protein conformation
from circular dichroismMethods Enzymol. 130, 208-269.
27. Hawes, J.W., Crabb, D.W., Chan, R.M., Rougraff, P.M., and Harris, R.A. (1995)
Chemical modification and site-directed mutagenesis studies of rat 3-hydroxyisobutyrate
dehydrogenase. Biochemistry 34, 4231-4237.
Page 19
Figure 1. Structural Features of Substrates and Products of Pyrroline Carboxylate
Reductase. Geometry optimization and calculation of torsional bond angles were performed
using the HyperChem software package employing the Polak-Ribiere (conjugate gradient)
method. Red indicates oxygen atoms; light blue, carbon atoms; dark blue, nitrogen atoms;
yellow, sulfur atoms; white, protons. Numbering of each of the rings (including the proline and
thiazolidine rings) corresponds to the conventional numbering of a pyrrole ring, with N as 1 and
the carbon adjacent to the carboxylate as 2.
Figure 2. Conversion of γ-glutamate semialdehyde to pyrroline carboxylate, and the
P5CR-catalyzed production of proline.
Figure 3. Inhibition of NADP+-dependent oxidation of thiaproline by L-glutamate and L-
proline. P5CR activity was measured exactly as described in Materials and Methods.
Glutamate and proline solutions were prepared in the same buffer as the enzyme assay (300 mM
Tris-HCl) and were adjusted to pH 7.2 directly before performing the assay. ■ indicates L-
proline; ♦ indicates L-glutamate.
Figure 4. Amino acid sequence homology between β-hydroxyacid dehydrogenases and
pyrroline carboxylate reductases. Sequences taken from NCBI databases were used for pair-
wise alignment to the rat hydroxyisobutyrate dehydrogenase sequence using either BLAST or
tBLASTn. Shown in black boxes are conserved residues, both identical and semiconserved,
which appear in at least 50% of all of the sequences. Arrows and text indicate functional
domains previously established for several of the β-hydroxyacid dehydrogenases. Blue boxes
indicate the motif GXXGXG in the catalytic domain. Red boxes indicate the invariant catalytic
lysine residue previously proposed in the β- hydroxyacid dehydrogenases. Stars indicate the
position of residues that were mutated in the E. coli P5CR during the present study. AE000357,
AE000157, and AE000394 represent loci in the E. coli genome coding for the corresponding
tartronate semialdehyde reductase sequences. HIBADH, hydroxyisobutyrate dehydrogenase;
RN, Rattus norvegicus, PA, Pseudomonas aeruginosa; AT, Arabidopsis thaliana; GM, Glycine
Page 20
max; PS, Pisum sativum; EC, Escherichia coli; BS, Bacillus subtilis; PA, Pseudomonas
aeruginosa; HS, Homo sapiens.
Figure 5. CD Spectral Comparison of Hydroxyisobutyrate Dehydrogenase and Pyrroline
Carboxylate Reductase. Rat β-hydroxyisobutyrate dehydrogenase (HIBADH) was purified as
previously described (21). Each protein was analyzed by CD spectrapolarimetry exactly as
described in Materials and Methods. Spectra shown are averaged from five successive
measurements.
Figure 6. Effect of Iodoacetamide and Tetranitromethane on Wild-Type P5CR. Treatment
with iodoacetamide and tetranitromethane and enzyme assay were performed exactly as
described in Materials and methods. IAA, iodoacetamide. TNM, tetranitromethane.
Table 1. Kinetic Parameters of Wild-Type and Mutant P5CRs.
Enzyme
Km NADP+ (mM)
Km T4C (mM)
Vmax (mmol/min/mg)
Kcat (Vmax/
nmole of enzyme)
Kcat/Km
NADP+
Kcat/Km
T4C
Wild-Type
0.18
0.30
1.70
11.97
66.510
39.90
Thr122Ala 0.24 3.73 0.67 1.74 7.27 0.47
Thr122Ser 0.67 2.53 1.11 3.71 5.54 1.47
Tyr35Phe 0.27 0.58 8.40 116.67 432.10 201.15
Tyr178Phe 0.33 0.91 4.61 80.88 245.08 88.87
Tyr201Phe 0.83 0.89 10.86 434.40 523.37 488.09